RE: Custom Dictionaries

[Sarah Carroll <sarah@xxxxxxxxx>]

Hi Stuart and all,

I guess that would be me and not the "dummy" Sarah :-) (although she's pretty much an expert too).
At 14:01 10/03/00 +0800, Stuart Burnfield wrote:
>Possibly translation software supports one or both of these
>ideas already -- I'm no expert. Sarah?

Interesting hypothesis. For a number of reasons, translators using the TRADOS Translation Solution (Translator's Workbench, MultiTerm and S-Tagger in this instance) work outside of FrameMaker, without access to the FrameMaker dictionaries. So for interactive translation, at first look, this might not be as beneficial as possible. For translators who work directly within FrameMaker I would imagine that ensuring that the correct terminology is used is a bit more difficult.

During interactive translation with Translator's Workbench (TWB), the translator also runs MultiTerm, which is the terminology management application used with TWB. It's pretty cool... Terms are entered into the MultiTerm database, along with their translations, and perhaps some contextual information. During translation, when a known term is found in a segment the term is isolated, highlighted and the information displayed in a special window in TWB. The translator can then choose to insert the translated term into the new translated segment (without having to type it), or to substitute it with something else if necessary and appropriate.

You are perfectly correct when you say:
<snip>
>restricted vocabulary of general English words, plus one or more
>sets of technical or jargon terms. The core terms are translated
>once for each target language, providing translators with a
>massive head-start for future jobs.
</snip>
A fundamental starting point for most translation is the generation of a glossary of terms and its translation. This is even more important with material originating within a Controlled Language environment. Where custom dictionaries could be of a great deal of assistance to those working with TWB and MultiTerm, could be in an export capability which would allow automated creation of a term base more easily. Most other translation memory systems include a terminology management application as well.

Where it begins to get tricky is the idea of "forbidden" phrases, constructs, words. At what point does the system tell the author that the text they have just inserted is forbidden? Possibly something along the lines of MS Word's wavy underlines on mis-spelled words (mis-spelt? sounds German). Anybody using a CL system which has such a function which works well? I'd be interested to hear.

There's absolutely no doubt, however, that the material which has been created using a CL system provides a significantly improved basis for translation using a translation memory system. Many people are looking at how to apply the technology to the authoring of the original material. Some interesting times ahead!

Best
Sarah


Please Note: New Telephone + Fax Number 
****************************************************
Sarah Carroll
Product Manager, TRADOS Ireland Ltd.
24 Longford Terrace, Monkstown, Co. Dublin, Ireland.
Tel: + 353 1 2020072
Fax: + 353 1 2302207
Mobile: + 353 (0)87 2401615
E-mail: mailto:sarah@trados.ie
Web: http://www.trados.com

****************************************************

** To unsubscribe, send a message to majordomo@omsys.com **
** with "unsubscribe framers" (no quotes) in the body.   **